Choosing Between Quantitative and Qualitative Evidence for a Community Profile
A student is writing a 1,500-word community profile of an Aboriginal-controlled health service in a regional town of about 12,000 people. She has three sources: (a) Census data showing that 6.8% of the town's population identified as Aboriginal or Torres Strait Islander, with a median age of 22 compared with 41 for the town overall; (b) the service's annual report, which states it delivered 9,400 episodes of care and employs 31 staff, 24 of them Aboriginal; and (c) two interviews with staff who say clients avoid the local hospital because of past treatment. The student's draft opens: 'The data prove that the Aboriginal population is young and that the service is trusted, while the hospital is not.' Her word limit is tight and she is considering cutting the interviews to make room for more statistics.
1. Evaluate the claim in the draft opening sentence, identifying which parts are supported, which are overstated, and why.
Read the worked answer
The sentence contains three claims of very different strength.
'Young' is well supported. The median ages (22 versus 41) come from a defined source and directly describe the population. Even here, the student should cite the Census year and collection district, because a town boundary and an Indigenous Area boundary rarely coincide.
'Trusted' is overstated. The annual report shows activity (9,400 episodes; 31 staff, 24 Aboriginal) and Aboriginal employment is meaningful evidence of community control and cultural safety. But service volume can reflect lack of alternatives as much as trust, and the report is produced by the organisation itself, so it is not independent evidence of client sentiment. The word 'prove' is the main problem: the data are consistent with trust but do not establish it.
'The hospital is not trusted' is the weakest claim. Two staff interviews are a small, purposive sample, and the interviewees are employees of a competing provider, which creates a conflict of interest. Their account is valuable as testimony about why some clients avoid the hospital, but it cannot support a general statement about the hospital's standing. The claim should be reworded as reported perception, attributed to named-role interviewees, and clearly limited.
Overall the sentence commits the error of converting descriptive statistics and a handful of interviews into proof of attitudes. Rewriting it as three separately hedged statements would fix most of the problem.
2. Should the student cut the interviews? Recommend a decision and justify it in terms of what each source type can and cannot establish.
Read the worked answer
She should not cut the interviews; she should cut or compress redundant statistics.
The sources answer different questions. Census data establish the size, age structure and geographic distribution of the population; they cannot explain behaviour or meaning. The annual report establishes service output and workforce composition; it cannot establish why clients choose a service. The interviews are the only source that speaks to reasons, experience and perception, which is precisely what a community profile needs if it is to go beyond description.
Removing them would leave the profile able to say that the population is young and the service is busy, but unable to say anything about access barriers, which is the most policy-relevant finding. It would also leave the hospital claim with no evidentiary basis at all, forcing its deletion rather than its qualification.
Where to save words: report only the Census figures actually used (perhaps three, not a full table), summarise the annual report in two sentences, and give the interviews roughly 150-200 words with clear attribution and a stated limitation. Keep a sentence acknowledging that two interviews are not representative and that further consultation would be needed.
The general principle: retain sources that supply a distinct kind of evidence, and cut duplication within a source type. Qualitative material is not decoration added after the 'real' data; here it carries the analytical weight.